Plasma HIV-1 RNA levels are the best predictors of long-term clinical outcome, mortality, and response to anti-retrovirals, but the prognostic value of these markers for neurocognitive outcomes is not known. If a reduction in systemic viral burden can reverse or stabilize neurocognitive impairment[minor cognitive/motor disorder (HIV-MCMD) and HIV dementia (HIV-D)] this would have tremendous treatment implications for the 20% of AIDS patients who will eventually develop HIV-D. Moreover, if the potential reversibility of neurocognitive impairment is mediated through a reduction in plasma viral load, then aggressive therapy of HIV-MCMD may be warranted. In both cross-sectional and longitudinal analyses, we will determine if plasma HIV-1 RNA levels are independently associated with specific neurological, neuropsychological, psychiatric and functional abnormalities in a cohort of 460 men and women with advanced HIV infection seen every 6 months at the University of Rochester, Columbia University and John Hopkins University. We will determine whether change in viral load is related to the onset of HIV-MCMD and HIV-D and transition among the stages of no impairment, HIV-MCMD, and HIV-D. Longitudinal analyses will include analyses of repeated outcomes (generalized estimating equations), time to event analyses (Cox models), and transition analyses (Markov chair analyses). In similar analyses, we will determine if plasma markers of immune activation (beta2-microglobulin, TNFalpha, TNFr2, sICAM- 1) are independently associated with neurocognitive impairment. Combining a carefully conducted natural history study with sensitive molecular assays will lead to a better understanding of how plasma viral load measures can be used to predict and monitor neurocognitive outcomes.